The Folk Song Index and Broadside Index are compiled by Steve Roud and are included on the Vaughan Williams Memorial Library (VWML) website with his permission. The two indexes share exactly the same field structure, although certain fields are more relevant for one type of material than another. The indexes are designed to be used in tandem or separately as the need arises and, on the VWML website, it is easy to search them one at a time or together.
The Folk Song Index seeks to provide details of English-language traditional songs which have been recorded in any medium - books, journals, newspapers, manuscript collections in public and private hands, published or unpublished sound recordings, videos, websites, and so on.
Entries in the index are made on a one-to-one basis. If a song sung by a particular singer appears on a record, it is given an entry in the index. If the same recording is issued on a different record, it gets another entry. If the song is then transcribed in a book, another entry is created. And so on. Cross references are given when possible, but users should be aware that the number of times a song appears in the index is not an exact indication of how many times it has been collected.
Roud numbers are assigned to songs to help with both identification and location and to counter the problems of multiple and variant titles. Thus, the song 'Seeds of Love' has been assigned the Roud Number three, and a search on this number will bring up all versions, whatever the title, wherever found.
The classification of folk songs is far from straightforward and a numbering scheme such as this is a blunt instrument designed to be practical rather than academically sound. It relies on the fact that in Anglo-American tradition the norm is a relative conservatism and that most song texts remain remarkably stable across both time and geographical space, which allows us to view them as 'the same song'. This does not work well across the board, however. There are some songs which are made up of 'floating verses', and there are certain subgenres (eg sea shanties and singing games) which are less well-served by this approach, but in the vast majority of cases the system works well enough to be useful.
The system is based on the comparison of texts; tunes are not included in this equation at all. For the same number to be assigned to two texts, it is not sufficient for them to share the same plot or characters; there must be strong similarities of wording. Something around 60 per cent of the wording needs to be the same for a confident match. But again, there are complications. Texts A and B may be borderline 'the same', and Texts B and C similarly connected. They would thus share the same number, even though A and C may not obviously qualify to be regarded as 'the same song'.
Even within this broad scheme, there is scope for different perspectives. When comparing versions, some indexers give more weight to history and development over time. Others, however, are more concerned with how the songs have turned out, and how they compare at the point of collection. One foregrounds the past, the other the present.
So, for example, the connections between Texts A and D might only be obvious once a common ancestor (or more than one) has been identified. The 'historical' perspective would regard this as sufficient evidence for regarding them as the same song, but the 'as collected' perspective would not.
One problem with the 'historical' approach is that it is sometimes difficult to be sure of the developmental lineage within a multi-faceted tradition when the evidence is necessarily partial and patchy. A problem with the 'as collected' approach, however, is that the texts being compared may be from very different times, and their degree of similarity thus obscured. In general, the 'historical' perspective is employed in the Folk Song Index.
There is no definitive answer in these cases. What suits one group of users will not suit another and, indeed, any individual user may well wish it to be one way on one occasion, and the other way when it suits.
The unavoidable fact that every decision helps some and hurts others can be clearly seen in two other examples. Earlier versions of the Index took more notice of function ad form of the songs. Thus, all 'alphabet songs' were given one number, regardless of whether they were the Sailors', Lumbermen, or others' versions. Following the logic of 60 per cent textual equivalence, these songs have now been separated, because it was really only the form, and the letters of the alphabet, that were the same. But people seeking Alphabet songs in general have had their task made much more difficult.
Similarly, some songs were previously grouped by function - all Wassailing songs, for example, were given one number, as were May Day songs, Pace Egg songs, and so on. As research identifies markedly different songs within these groups, they will be given new numbers.
It follows that as more versions and more information come to light, numbers have to be revised. Sometimes a song is assigned a new number as previously unnoticed connections are made, or it is thought best to split a group of songs into two. A list of changed numbers is maintained, which will eventually be available online. No defunct number is ever used again for a different song.
Numbers are assigned sequentially and have no intrinsic meaning. The next song discovered gets the next available number. The Roud numbers in the Folk Song Index are plain numerical, starting from one. Numbers in the Broadside Index start with 'V' (see below).
Once a song has been identified as qualifying as a 'folk song', entries for non-traditional versions (eg broadside printings) are copied from the Broadside Index into the Folk Song Index, in order to ensure that the latter remains one-stop-shop for details of traditional songs.
Originally designed as an adjunct to the Folk Song Index (see above) to aid with historical research in that field, the Broadside Index has now achieved a life, and importance, of its own.
The bulk of the entries in the Broadside Index comprises references to songs published on printed street literature trade (late 16th to late 19th centuries) in the form of broadsides, chapbooks, and cheap songsters. But coverage has been extended to include a much wider range of popular and vernacular songs including, in particular, eighteenth century songbooks and nineteenth century music hall publications. Originally envisaged as covering Britain and Ireland, material from across the English-speaking world is now included. The cut-off date is about 1920.
Roud numbers are assigned to songs to help with both identification and location and to counter the problems of multiple and variant titles. Numbers in the Broadside Index, for songs which have not been identified as 'folk songs', start with 'V', and are assigned sequentially, ie the next song discovered gets the next available number. The number has no intrinsic meaning (see above for further discussion of Roud numbers).
As detailed in the guide to the fields, where possible live links are included within the Index. Broadsides which are available online, the appropriate URL is included in the URL field. Where the printer/publisher of an item has been included in the Printers' Register database, a link is included in the PRINTERID field, and further details of the source of a particular item can be found through the link to the Bibliogaphy in the ROUDBIB field.